Socioeconomic Drivers of Suicide Rates across European Countries: A Bayesian Model Averaging

Abstract

This article aims to examine socioeconomic determinants of suicide rates across European countries. Using Bayesian Model Averaging and panel data specification with fixed effects on a sample of 25 European countries and two subsamples between 2001 and 2019, a data-driven model was estimated. Empirical findings for the whole sample suggested GDP per capita as a regressor with very strong posterior inclusion probabilities and negative effects of GDP per capita on suicide rates. Marriage rates appeared with positive inclusion probabilities and effects from marriage rates to suicide rates were negative. Share of population ages below fifteen years and divorce rates were identified as regressors with weak inclusion probabilities. The effects from a share of the population aged below fifteen years to suicide rates were negative while effects from divorce rates to suicide rates were positive. Empirical findings for a subsample of post-socialist countries revealed divorce rates as the most prominent regressor explaining suicide rates in post-socialist countries with very strong posterior inclusion probabilities. Female participation in the labour force and share of the urban population appeared as regressors with strong prior inclusion probabilities. The effects of female participation in the labour force on suicide rates were negative as well as the effects of urbanization on suicide rates. GDP per capita appeared as a regressor with positive inclusion probabilities and the effects of GDP per capita on suicide rates were negative. For a subsample of non-socialist countries, empirical findings suggested a share of the urban population and female participation in the labour force as regressors with very strong posterior inclusion probabilities. Unlike the estimates for the subsample of post-socialist countries, effects from participation in the labour force to suicide rates as well as from urbanization to higher suicide rates were positive.

Keywords

• suicide
• socioeconomic factors
• data-driven model
• bayesian model averaging